How do you keep a balance between AI, data, and ethics?

Tim Hunter is a Senior Artificial Intelligence Specialist at ABN AMRO. Read about his insights on AI, data, and ethics in this article.

The digitization we are making as a bank goes hand in hand with the use of artificial intelligence. Algorithms help us personalize our services, optimize internal processes, and prevent fraud and money laundering

Creating the future
Tim Hunter is an Artificial Intelligence (AI) Specialist within ABN AMRO's Group Innovation department. He studied AI and Machine Learning at Berkley University in California. He started his career as a software engineer at Databricks. Databricks is an Open Source platform that simplifies data architectures by eliminating the data silos that traditionally separate analytics, data science, and machine learning.

"I was asked to help set up a new branch office in Amsterdam. That is how I got in touch with ABN AMRO. I made the switch because, at ABN, I would get the opportunity to radically change the company using AI technology. At Group Innovation, my colleagues and I guide the search on how to apply the possibilities offered by AI. In close cooperation with the business, we are enthusiastically working to create the bank of the future."

One of the biggest challenges I face is to implement AI in a responsible, transparent, and explainable way. The following questions are important:
-Is that client eligible for a loan?
-Do we see fraudulent behavior?
-Is this new product interesting for a specific client?

These questions can be answered with the help of AI. However, are these the right or correct answers, and more importantly, are we, as a bank, allowed to use the data that will provide us with these answers?

For instance, some patterns in payment behavior are a predictor of financial trouble. Is it prudent, responsible, or in line with regulations to sort based on these patterns? Even though laws and regulations provide a direction, ABN AMRO operates from the principle that AI can only be applied if the results are retraceable and fit our standards and values. That is why we do an ethical assessment together with a large group of colleagues for every AI project.''

Financial crime
"A lot of the algorithms we use are Open Source. The important thing is to know how to train these algorithms afterwards. To train them, it is crucial to have available data. Therefore, AI is particularly applicable in the field of fraud prevention. That's the area that I am focusing on right now. Based on historical data, we can discover patterns and signals that help us detect suspicious cases.

That is how AI makes it possible to identify transaction groups that deviate from normal patterns. After detecting them, these groups of transactions are checked by a fraud expert. The fraud expert provides context to this 'unusual behavior', and decides if it is actually a case of fraud. Mainly, AI provides more efficiency. It helps the fraud expert to focus on the cases that require special attention. For me, I am very happy to contribute to that."

Click here for current job openings.